ZBrushCentral

Will a new processor really make any difference?

Community
#1

Hey

My current PC specs are.

Intel i7-2600k 3.4GHz
16 G ram, I think its DDR3 1600Mhz but im not certain.
Geforce GTX 670
Im running a decent SSD aswell.

Those are what matters in this case.

My freelance career has moved to producing 100 mill poly sculpts at the very top end and of course my PC is chugging some what. I normally have 15 / 20 different meshs in one scene and working on one in isolation is fine, each mesh is normally no higher than 20 mill tris.

When it comes to seeing the entire piece as one, all meshs visible, I do have to rotate and move at one of the lowest 3 sub div levels and only push it up to max for still viewing. It normally takes roughly 5 minutes for everything to catch up each time I press a button or move the mouse at that poly count.

I know ZBrush only uses 4g of ram so thats not really an issue, as is the graphics card.

If I upgraded my processor significantly, to a current high end with of course a new mobo, would that really make a noticable difference? I dont know.

Hope someone can advise.

Many Thanks

#2

As of 4R7, it can use more than 4GB actually. By using the 64bit version that’s included in the install directory.

A better CPU will help, as will higher memory bandwidth/speed. Go with one of the quad-channel memory setups that the LGA 2011 socket CPUs provide.

But also try the 64bit ZBrush first… Your mesh needing to be compacted and written to disk could likely be why it takes so long to do things. 64bit will open up a lot more memory for your models.

#3

Sorry, apologies… I am already using the 64 bit version of ZBrush. As soon as you mentioned that I realised my mistake. Why that slipped my mind when I posted this morning I dont know. All of my original post still stands, just with a 64 bit version :slight_smile:

Knowing that, would one of the new CPUs you talked about combined with equivalent Ram give me a noticeable / worth while for cost improvement in handling up to 100 mill scenes?

Many Thanks

#4

Ah ok hehe, it happens :stuck_out_tongue:

Well, upgrading to say a 4790k would be a bit of a bump, but not really worth the upgrade in my mind, you might get a 10-15% performance increase at best.

A 5820k would be a better upgrade, as it has 6 cores (12 threads) and quad channel, though I’ve read that ZBrush only uses 4 cores at best? But those were older posts, maybe 4R7 fixed that? Because when I have my Task Manager open, it shows all 8 threads being used evenly…

Also, if you’re already populating 4 channels with RAM on your system, then you could reuse those in your quad channel motherboard, saving some money (though you might want to splurge for faster RAM depending on what you have right now)

edit: Oh, also make sure you’ve increase your “mem” preferences in ZBrush to use more RAM and to wait for a larger size before compacting memory.

#5

Thanks for the reply.

Yes I agree, it would have to be a major CPU upgrade but I think it would make a noticeable, not massive but noticeable difference. Ill have to look into prices and as you say, find out for sure what my current RAM is.

Good tip about Preferences - Mem… I had forgotten about that.

Thanks!

#6

If you’re talking about single subtools with meshes that approach the 100 million poly mark, nothing will help much with that. That’s simply near the upper (and not recommended) limit of Zbrush’s per subtool limit, and the performance and stability with those tools is likely to be shaky on most hardware. Examining your workflow for better polygon efficiency would pay greater dividends than better hardware in this case.

If you’re talking about files that are cumulatively 100 million polygons across many subtools, more/faster Ram could likely give you better breathing room. You could upgrade your CPU, but the price you’d have to pay to see a really noticeable upgrade would be significant.

A CPU might give a bit more speed to operational crunch and brush performance, but wont help with max file size.

#7

I am talking about multiple meshs in one scene all combined totaling 100 mill. Each mesh is no higher than 20 tops.

Faster RAM is the easiest and cheapest option, I’ll need to look into what my mobo can support.

Yeah the CPU upgrade would have to be significant to be notable. Operational performance improvements is what I’m after… definatly don’t need more polys hehe.

Many Thanks